cxx_assign_const produces wrong diffs

[Reen]

Hi,

thanks for this great tool.
I'm trying to use it on my code and I get patches like this from the cxx_assign_const mutator:

survivedfoo_hpp-cxx_assign_const-L81-C34.patch

--- a/foo.hpp 0
+++ b/foo.hpp 0
@@ -81,1 +81,1 @@
-        _inv_rotation_matrix[0U] = matrix[3U] / det;
+        _inv_rotation_matrix[0U] 42 matrix[3U] / det;
--
Build Date: 09 May 2023
Commit: 5965c74b
LLVM Version: 14.0.0
Mull Version: 0.20.0
URL: https://github.com/mull-project/mull

I assume, that it mutated the matrix[3U] into 42?

Environment

Ubuntu 22.04
Clang 14
Built from source
Commit: 5965c74

[AlexDenisov]

Hi @Reen, apologies for the late reply.

We are using debug information for such patches/replacements, I think in this case the debug info is pointing to = so I think the mutant was of this form:

_inv_rotation_matrix[0U] = 42;

[AlexDenisov]

Hi @Reen, I've fixed the issue here, the right patch for your case should produce _inv_rotation_matrix[0U] = 42; matrix[3U] / det; as we do not eliminate the right-hand side of the assignment.
I'll cut the new release shortly, please to not hesitate to reopen the issue if the problems persists or you find another edge case not properly handled.

[AlexDenisov]

Here we go https://github.com/mull-project/mull/releases/tag/0.21.0